 Thank you. That was a wonderful talk, Jonathan, and a great introduction to the one I'm going to give. Mine will not be nearly so coherent, but I hope it will be thought provoking. I have no specific disclosures. I do want to acknowledge in my multi-center studies, there's a cast of characters that's enormous, but where I have recently landed in the Longwood area, there are some particular individuals who I owe great debts of gratitude for, for the collaborations that we're pursuing and what they've taught me. Now in this conference, it sounds like what we'd like to be, see ourselves as, is archers. This archer theme struck me, so we'd like to see ourselves as glamorous archers who are zeroing in with great romantic accuracy on the target, whereas we are actually struggling. And I'm afraid that what I have to say is not going to help because I do think that the issue, the fundamental issue, is when we have all this information, as Jonathan pointed out, I like the term incidental. What do we have to look for and what do we have to disclose? I do think that we need to be clear in our minds whether we're talking about disclosing to an individual, to a consumer, to a layperson, or we're disclosing to a medical doctor, and you'll see what I mean in a moment. I also do think, as I alluded to in my comment before, that there's some low-hanging fruit to be gathered here. I love this diagram that Dr. Manolio put in her paper, and it reminds us that we can really talk about some of the rare, highly penetrant, tightly linked disorders before we have to talk about more of common variants which are more controversial. Now in answer to the question, what do patients or consumers, what do they want to be disclosed, I think the answer is very clear. It varies. And it varies, let me just give you a couple of examples. It varies from George Church's PGP project, which is this open access to data where people are volunteering to have their entire sequence placed on the Internet along with their health records, multiple traits, cell lines used for all sorts of research, and this IRB has approved it for 100,000 volunteers, and they've already got 16,000 volunteers, of which 2,000 and some have already put past a genetics entrance exam in order to demonstrate that they can handle the information, and over 1,000 have put their medical records online, and they have 500 genomes in the pipeline. I take this at one extreme. We also have the example of direct-to-consumer testing, which I think is, of course, controversial in many arenas, but for me it's this incredible natural experiment in incidental findings. I mean of the over 100,000 individuals who've ordered their own genetic markers through various companies, most of them had no idea what it was they were going to get. So most of them, they might have had one or two things they were interested in, they might have been interested in ancestry, but when they actually got it, there they were faced with either findings that were not particularly impressive or findings that were actually new and incidental to them. We have a study called PGEN that we have been funded by NHGRI, where we're going to be doing the first, before and after interrogation of customers and two companies with their genetic data in hand that they allow us to have in order to try to understand what it is they thought they were going to get, what is it they actually got, and what is it that they understood or misunderstood about what they found out. And this is particularly relevant now that some companies have started to give actual Mendelian disorders as well when company is now giving BRCA variants and so the door is opening to exome and to sequencing from these companies another reason why we need to get our act together or else I think we will be somewhat bypassed by the commercial sector. Now can we even define clinical action ability? I'm with David in really arguing against this term has much meaning and I think the answer is probably not. I hope so anyway because I'm debating Jim Evans tomorrow afternoon at the primer meeting and he's going to argue that there is such a thing as meaningful action ability and I'm going to argue that there's not. But this slide which is courtesy of David Kaufman and some focus groups that he has run, this is not my work. I think really puts it very nicely. So he describes a narrow definition of clinical utility which is the information can really help you treat or avoid disease, a much broader definition of clinical utility where you can motivate some behavioral changes, maybe it's one of a number of things that help you motivate change. You can just learn more about your own condition. You can monitor research and progress, maybe you donate to progress. You can participate in related research. You could be useful in the future with things. But then it goes on down to what we and others have been calling personal utility, empowerment, giving a feeling of control, responding to your curiosity, feeling respected by the researchers, involving your family, helping you to plan or live more fully and then other sorts of reasons that are even beyond those. Feeling a sense of ownership, your results belong to you, participants simply want to know what it is you know about them or feeling like the results are kind of compensation for participating in the research. Now, in these comments, I'm going to switch back and forth between clinical and research only because our thinking about incidental findings has been more advanced right now in terms of clinical, I mean in terms of research issues than it has been in terms of clinical issues. So as many of you know, I've been leading the reveal study which has been funded for about 11 years now in which we've used the paradigm of ApoE disclosure for Alzheimer's disease to drill down and learn a lot about what it means to disclose susceptibility information. And Mark, I certainly hope there are going to be categories of variants and that this can serve as one of the sentinel categories. Otherwise, I've spent 10 years working on one variant and that's just a long time if it's not in some fashion generalizable to the rest. And what we've seen is very interesting, we didn't even know if anybody was going to want this information when we clearly explained to them how probabilistic it really was. But in fact, 24% of the people that we specifically and systematically approached actually put out their arms and volunteered for the study. And of course, it was a much higher percentage of the people who sought us out, people who really wanted to know. But we were also reassured by the number of people who did not want to participate because people were understanding that they had a choice and making the choice that they didn't want to know this probabilistic information. When people did, of course, it's the ones who signed up, when they actually did get their results, they were very happy to get their results. And as most of you know, the flagship paper actually showed that there was not really any increased anxiety or distress associated with the disclosure of ApoE in the people who wanted to learn it. But it showed that they were pretty happy with this and we'll drill down on that in a minute. We also had a measure of their value that they ascribed to this experience called willingness to pay. And it's just an exercise where you sort of ask people, what would you pay for this? And we asked them after they had gotten it, after they had learned what they had. And it turned out that people were not actually learning whether or not you were E4 positive did not influence the value that you ascribed to this probabilistic variant. The only two things that were associated with what, how much you said you would pay were your income level. So if you've got some discretionary money, this is one way you might choose to use it. And your increased desire to know your AD stats. So people who really wanna know and people who have money are gonna find ways to find out. And if they can't do it in the medical care system, I suspect they will go around the medical care system. Now we also feel that we challenge this prevailing notion that when people receive probabilistic genetic information, they won't do anything with it. We've found early on that people tried to do things even when we told them there were no proven treatments for Alzheimer's disease. So they're more likely to change their vitamins, exercise, and medications. We found there was a dark side to that too, because we found that they were more likely to order unregulated supplements off the internet. And when there was no regulated treatment for preventing Alzheimer's disease. And we were the first study to show that people were more likely to purchase long term care insurance if they learned they were E4 positive. Something that gave the long term care insurance adjusters conniption fits. But we also found some curious things about liking and understanding their information. This was a paper that I was particularly proud of the title. Because the title of this paper is, I know what you told me, but this is what I think. And it sort of reminds me of my teenage son. But in this paper, we've discovered that people who accurately remembered everything we told them about their risk information. We tested them, and they accurately remembered it. Turned around and said, I know what you told me, and I agree with my risk that you told me, or I know what you told me, but I believe my risk is actually lower or higher. So people seemed, and this was correlated with their risk as they walked into the study. So people have an anchoring effect that's framed by many different things. And their family experience, their perception of the illness. And this is something that they walk in with. And it may not be so fragile as we have once imagined for us with genetic information to actually move that dial one way or another. We also looked in depth at people 12 months after they learned their genetic information and their risk information. And we found that in fact, the pros that they had selected at the beginning of the study, that they were interested about information for preventive measures, that they were gonna make arrangements for long term care, and so forth, actually degraded over time. And that was an interesting phenomenon. They became a little bit less enamored of the pros. And they also became a little bit more enamored of the cons. So there was a kind of mediating effect in which they realized that the information they had had some limited value. And we think that's a good thing. That's a natural thing in terms of information. In terms of the comments about social networking, people tell this information to all sorts of people. We were amazed. People, we've warned people, don't tell anybody, this is confidential. People could put it in your medical record. They told all sorts of people. And 12% of them told their health professional, even though we'd explicitly warned them in the consent not to do so. We've put over 1,000 people through four different randomized clinical trials with ApoE and Alzheimer's disease risks. The third one, which we're just analyzing the data now, was actually specifically about incidental results. We surprised one arm of the study with the fact that ApoE4 is also a risk factor for cardiovascular disease. Don't ask me how we did that with the IRB. It's a different discussion, but we did. And it was quite interesting. Not only did people not have greater or lesser distress, but they had a rational response to the incidental finding. Whereas when they hadn't learned about the additional risk of cardiovascular disease, they did not have much of an increase in terms of their exercise. Remember, we demonstrated that they had a little bit before between E4 negatives and E4 positives. When they were in the arm that were told about E4 and the risk of cardiovascular disease, they increased their exercise, or at least reported to us that they did so. We all know that real behavior is hard to change. And this is a rational response to incidental findings, and they were not dissatisfied to receive that information. I've had the opportunity to work in a small way with the gene partnership program, which I'm delighted to be affiliated with. And you've heard a little bit about that and the sort of idea that people can actually exert a preference and tune in or out the kind of information that they want. And working with Ingrid Holm, who's received several grants from the Genome Institute vis-a-vis understanding preference setting, we have some really interesting preliminary data that I thought I'd share with you. First of all, we recapitulated what Kathy Hudson and others have found that people really do want the results of genetic research. And they want it in a large majority of cases, and they say that they are more likely to enroll in research if they get it. Again, we're cognizant here of the difference between research and clinical. I'm going to get back to clinical, but this recapitulates what we found before. But look at this. If you separate people into those who basically say they want all the results and separately, basically people who say, I don't want all the results, I want to choose my results. Remember, a majority were in this group. And then when you drill down and say to people, do you want results whether the information is well-established or not well-established? They say, yeah, I want it all. Do you want the results if you're a lot more likely to get the disease or a little more likely to get the disease? They say, yeah, I want the results. Do you want it preventable or not preventable? Fatal, not fatal. Childhood onset, adult onset, pretty much, 75% of them across the board, sometimes 80%, 90%, regardless of the condition, say, yeah, I want the results. If we're going to go with preference, we're going to be giving out a lot of information. I actually have a theory about this, which I'm calling the diagnostic misconception. I wish I'd made up that term, but I didn't. And I hypothesized in this survey that people who said they wanted to receive all this information were more likely to believe that that information would directly help them. In other words, if you're in a research study and you're going to get information back about yourself, you actually don't believe the researchers would be so foolish as to give you back information that didn't make any sense. So I think there is an assumption underlying the very concept of giving back information that says, well, you're not an idiot. You're going to give me back information that means something. And I call this the diagnostic misconception. Let's switch to clinicians. We just talked about the participant or patient perspective. What about clinicians? Well, I also think it varies. And I'm going to show you for the first time some results that we've never shown in public before. I queried a fairly arbitrary list of individuals, several of whom are sitting in this room, with a spreadsheet that Howard Jacob and I designed in which we basically listed the top 88 conditions in gene tests. And we said, please pick the ones that you would feel you would recommend to a molecular laboratory to return to the referring physician that you feel should be returned. We deliberately worded it a little ambiguously. And take a look at this. If in terms of returning to adult patients or to the physician of adult patients, there was 100% concordance for 21 conditions. I'm going to show you what those are. But there was 80% concordance for 65 conditions in adults and just as many conditions in children. And once you got up to a 70% concordance among these 15 experts, there was like 90 of the 99 conditions they recommended to be returned back. So now this is part of the low-hanging fruit I'm talking about. These are gene tests. They're generally Mendelian conditions. And these are the 21 conditions that these 15 somewhat arbitrarily selected experts actually all said should go back to adults if there was a known pathogenic variant. And the ones with an asterisk, they actually said should go back to children too. And the last couple slides, I just want to show you the differences in patterns because if we think we're going to get in a room and come to a consensus about how this should be done, I show you this slide to caution us. Here's, these are just three of the individuals. I don't even know who they are. But this individual basically created a pattern where they wanted to return about 40 of the conditions if it was a known pathogenic mutation. About 40 of the conditions if it was a truncating mutation suggestive of pathogenicity and hardly any of the mutations if it was a missense mutation that was predicted in silico to be pathogenic. This individual had a very different pattern of treating those three things and this individual basically would return them all regardless of the type of variant. Same kind of variation occurred when dealing with children. So you can see in this quick overview of data that I had either generated or had access to that there's a wide variety of ways in which patients or research participants for this discussion I'm lumping them somewhat together feel about return of results and how clinicians or researchers feel about giving them back. This is only gonna add to the difficulty we have in trying to come to consensus. But I will say that I believe the reveal study and other people studies so far have shown that there is very little harm in disclosing information back to individuals and that the myth of information as harmful may be doing quite a bit to retard the dissemination of information in genetics and we should perhaps take a new look at that as one of the downsides of genetic exceptionalism. Thank you very much. For this talk, so I'll lead off. So I think in your use of the term harm you're probably looking at it from the perspective of what a patient would perceive as being harmful. One might argue if you were a payer and you looked at behaviors, some behaviors which may in fact incur costs without necessarily an attendant benefit that that could represent harm and one of your colleagues Zach Kahani has talked about the impact of the incidental loam about activities that take place when something is found that really probably should be ignored. So could you comment a bit on that? Well I absolutely agree. I think harm to the individual versus harm to society, harm to the medical care system, waste to the medical care system, that's a different issue. Now in one of the grants that we'll be starting in the next month or so, actually in the next week or so, we have built into this to look at the outcomes associated with return of incidental information and sequencing including some measures of economic cost. Though in the timeframe that we can do the grant I'm not sure that it's actually fair. Both for the potential health savings and the potential health costs you probably need a longer timeframe. But I absolutely agree with you and that's why I raised Mark that comment before about sort of looking at it two ways in terms of what we should imagine regulating in terms of information being very different and what we might imagine regulating in terms of reimbursement. Erin? Robert in your survey of your group of experts did you address the logistics of reporting out 80, 90 conditions because although you're right there's evidence that there's not a lot of harm to it. I think one of the thoughts for clinicians is logistically how is that even feasible and what are the additional harms when you're trying to report out 80 diseases at once what gets missed and just how is that accomplished? No, we remained artfully vague about a lot of things. We sort of assumed that sequencing was perfect. We assumed that you could even find chromosomal abnormalities, indels and so forth. We stipulated all this as sort of a given. But do remember that these are all individually extremely rare except for the heterozygous recessives that others have talked about. So you would presumably have a very small set maybe less than 12, even if you counted carrier states that you would give back to any given individual. Right, but that assumes a very, you know, using maybe a curation set like what Dr. Berg proposed which is very, very narrow and may not be ultimately what gets reported out when you consider pharmacogenomics and common complex diseases and things like that. Right, my point is more of it, we just have to start somewhere. And it seems that the items that are tightly bound to disease that we think actually might be meaningful in what is essentially a public health screening context would be the logical place to start. And I think it corresponds to some degree to the Ben 1 and Ben 2A, things that the group has reported there. I'll take one more question from Paul and then we'll go to the next one. Robert, Paul Billings from Life Technologies. I'm gonna make a kind of a Bruce Blumberg first comment which is that your 15 physicians are probably not that demographically diverse and they're kind of a narrow group. But I'm curious, did any of them talk about the existence of let's say an orthogonal test or a confirmatory test as helping them gate their activity along this one? Several of them are in the room and can respond but I did not hear that. I think several had spent several hours on this, struggled with it, but in order to make it meaningful it was almost no discussion. It was done individually. There was no consensus. There was no telephone call. Everybody had to make their own forced decision yes or no. Thank you very much.